Skip to content

Requirements

Before installing and using the MemVerge Transparent Checkpoint Operator, ensure that your Kubernetes environment meets the following system and software requirements.

Kubernetes Cluster

  • Version: Kubernetes v1.28 or later.
  • Role: Access to the Kubernetes cluster with the cluster-admin role is required for installing and configuring the operator.
  • Distributions: The following Kubernetes distributions are supported:
  • Vanilla Kubernetes
  • Rancher Kubernetes Engine 2 (RKE2)
  • K3s

Container Runtime Interface (CRI)

The MemVerge Transparent Checkpoint Operator is compatible with the following CRI runtimes:

  • Containerd: Version 1.7 or later.
  • CRI-O: Version 1.28 or later.

Other CRI runtimes are currently not supported.

Storage

  • A default StorageClass must be configured in the Kubernetes cluster.
  • The StorageClass must support dynamic provisioning of Persistent Volumes.
  • The StorageClass must support moving Persistent Volumes from one node to another. This is essential for the checkpoint and restore functionality.

GPU Drivers

  • NVidia Server Drivers
  • Driver v550.x.y or later: Supports 1 GPU per worker node
  • Driver v575.x.y or later: Supports multiple GPUs per worker node

Command-Line Tools

  • kubectl: Version 1.28 or later, properly configured to communicate with your Kubernetes cluster.
  • Helm: Helm package manager version 3.14 or later.

Networking

  • The Kubernetes cluster must have network connectivity to the GitHub Container Registry (ghcr.io/memverge) to download the MemVerge Transparent Checkpoint Operator container images.

GitHub Access Token

  • A personal access token from the mv-customer-support GitHub account is required to download the MemVerge Helm chart and container images. Contact MemVerge Customer Support at support@memverge.com to obtain this token.

Optional Requirements

Nvidia GPU Operator (for GPU Checkpointing)

  • If you plan to use the MemVerge Transparent Checkpoint Operator with GPU-enabled workloads, you need to install the Nvidia GPU Operator.
  • The Nvidia GPU Operator must be configured to enable Kubernetes's native Container Device Interface (CDI) mode. See the Installation Guide for specific configuration instructions.